Authors:
Gergely Varga
;
Sara Toth
;
Christopher R. Iacovella
;
Janos Sallai
;
Peter Volgyesi
;
Akos Ledeczi
;
Gabor Karsai
and
Peter T. Cummings
Affiliation:
Vanderbilt University, United States
Keyword(s):
Design, Experimentation, Metaprogramming, Web-based Development, Software Synthesis, Scientific Computing, Molecular Dynamics Simulation.
Abstract:
Molecular dynamics simulators are indispensable tools in the arsenal of chemical engineers and material scientists. However, they are often difficult to use and require programming skills as well as deep knowledge of both the given scientific domain and the simulation software itself. In this paper, we describe a metaprogramming approach where simulator experts can create a library of simulation components and templates of frequently used simulations. Domain experts, in turn, can build and customize their own simulations and the required input for the various supported simulators is automatically synthesized. The web-based environment also supports setting up a suite of simulation jobs, for example, to carry out automated parameter optimization, via a visual programming environment. The entire simulation setup – including the various parameters, the version of tools utilized and the results – is stored in a database to support searching and browsing of existing simulation outputs and
facilitating the reproducibility of scientific results.
(More)